From owner-freebsd-current Fri Jan 26 10:53:04 1996 Return-Path: owner-current Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id KAA17745 for current-outgoing; Fri, 26 Jan 1996 10:53:04 -0800 (PST) Received: from phaeton.artisoft.com (phaeton.Artisoft.COM [198.17.250.211]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id KAA17740 Fri, 26 Jan 1996 10:53:02 -0800 (PST) Received: (from terry@localhost) by phaeton.artisoft.com (8.6.11/8.6.9) id LAA04990; Fri, 26 Jan 1996 11:51:24 -0700 From: Terry Lambert Message-Id: <199601261851.LAA04990@phaeton.artisoft.com> Subject: Re: Good news -- pipe stuff To: dyson@freefall.freebsd.org (John Dyson) Date: Fri, 26 Jan 1996 11:51:24 -0700 (MST) Cc: current@freefall.freebsd.org In-Reply-To: <199601261541.HAA09194@freefall.freebsd.org> from "John Dyson" at Jan 26, 96 07:41:52 am X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-current@FreeBSD.ORG Precedence: bulk > Pipe latency is about 2x the speed, pipe bandwidth is about 50% faster, > and I am still evaluating the context switch -- but it is faster. Before > I include the changes, I plan to add more intelligent memory management > of pipes (including making them pageable -- instead of tieing up wired > memory), and page-flipping. Page-flipping should add about 50% more onto > the bandwidth. Good show on the pipe code! I read the Lai/Baker paper: in "Figure 1", FreeBSD kicks some serious butt on context switches -- it appears to be both flat and linear past 200 processes (the limit of the graph in the figure). I would be loathe to screw with what is essentially "perfection", IMO. Solaris is on the order of 3 times slower until 30 processes, and jumps to 6-9 times slower based on algorithm (the worst algorithm being the one they must use in the SMP case). Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.